[Mini-VLLM] Remove LoRA, plugins, speech-to-text and scale-out surfaces

Production runs one static model behind the OpenAI-compatible API. Every
adapter path and non-OpenAI HTTP surface is unreachable there.

- Delete the LoRA stack: vllm/lora, LoRARequest on the engine wire, the
  scheduler, worker, cudagraph and MoE plumbing, LoRA metrics, the lora
  CLI group, and the LoRA tests.
- Drop the plugin loader in vllm/plugins. Platform resolution is now
  CUDA-only and the general/stat-logger plugin hooks are gone.
- Delete the speech-to-text and realtime entrypoints, the scale-out
  render routes, and vllm/config/speech_to_text.py. GenerationTask is
  generate-only.
- Cut DecodeBenchConnector and the disagg-prefill doc and image. The
  Multi, Offloading and Example connectors stay; production uses the
  CPU KV offload tier.
- Delete the ROCm-only fused KDA chunk kernel (2346 lines) with its CMake
  arch gate, declarations and op registrations. The CUDA build never
  defined the guard and no Python caller existed.
- Delete the orphan Mistral and Pixtral chat fixtures and examples/.
- Fix _check_model: removing the LoRA early return made every request
  fail with "The model ... does not exist. Valid aliases: ...".

Verified on an RTX 3090: import sweep reports 1242 modules with 7 known
optional-dependency failures; a live server answers chat, completions,
responses, tokenize and detokenize with zero log errors and 404 for the
removed surfaces; collection yields 8558 tests with the 3 pre-existing
errors; ruff adds no new finding; the incremental _C build succeeds.

Co-authored-by: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
This commit is contained in:
2026-10-19 21:06:19 +07:00
co-authored by Hinata
parent 5f3c7bd8ab
commit 5aa2ac64f8
269 changed files with 417 additions and 29932 deletions
-28
View File
@@ -1,28 +0,0 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import json
import pytest
from tests.plugins_tests.test_oot_registration_online import (
run_and_test_dummy_opt_api_server,
)
def test_distributed_oot(dummy_opt_path: str):
run_and_test_dummy_opt_api_server(dummy_opt_path, tp=2)
@pytest.mark.parametrize("enforce_eager", [True, False])
def test_distributed_oot_pp_raw_input_tokens(dummy_opt_path: str, enforce_eager: bool):
# The model asserts that input_ids reaches every pipeline stage.
overrides = {"architectures": ["MyOPTRawInputTokensForCausalLM"]}
extra_args = ["-pp", "2", "--hf-overrides", json.dumps(overrides)]
if enforce_eager:
extra_args.append("--enforce-eager")
run_and_test_dummy_opt_api_server(
dummy_opt_path,
extra_args=extra_args,
env_dict={"VLLM_USE_V2_MODEL_RUNNER": "1"},
)
-10
View File
@@ -24,9 +24,7 @@ class _LegacyBlockStored(
parent_block_hash: bytes | None
token_ids: list[int]
block_size: int
lora_id: int | None
medium: str | None
lora_name: str | None
extra_keys: list[tuple[Any, ...] | None] | None = None
group_idx: int | None = None
kv_cache_spec_kind: str | None = None
@@ -58,9 +56,7 @@ class _PreSessionBlockStored(
parent_block_hash: bytes | None
token_ids: list[int]
block_size: int
lora_id: int | None
medium: str | None
lora_name: str | None
extra_keys: list[tuple[Any, ...] | None] | None = None
group_idx: int | None = None
kv_cache_spec_kind: str | None = None
@@ -79,9 +75,7 @@ def _make_block_stored(
parent_block_hash=None,
token_ids=[1, 2, 3, 4],
block_size=4,
lora_id=None,
medium="GPU",
lora_name=None,
group_idx=group_idx,
kv_cache_spec_sliding_window=kv_cache_spec_sliding_window,
locality=locality,
@@ -183,9 +177,7 @@ def test_block_stored_locality_is_wire_compatible():
parent_block_hash=None,
token_ids=[1, 2, 3, 4],
block_size=4,
lora_id=None,
medium="GPU",
lora_name=None,
group_idx=2,
kv_cache_spec_sliding_window=128,
)
@@ -221,9 +213,7 @@ def test_block_stored_session_id_is_wire_compatible():
parent_block_hash=None,
token_ids=[1, 2, 3, 4],
block_size=4,
lora_id=None,
medium="GPU",
lora_name=None,
group_idx=2,
kv_cache_spec_sliding_window=128,
locality="LOCAL",